Age | Commit message (Collapse) | Author |
|
am: 06e002d2cd
Change-Id: I168636930a95e3a11458537c2535b195571f9e6c
|
|
Bug: 32721372
BZ: IMINAN-51199
During the decoding process, there're >= 2 threads (such as ISV and
MediaCodec) which may share the same gralloc buffer
(obj_surface->share_info), although their allocated surfaces belong
to each corresponding decoding context.
So fixed the random issue by registering (referencing) the gralloc
buffer before locking it (for read/write), to avoid its potential
free by other processes/threads. And the gralloc HAL would
defer to free the buffer once all of the registered processes or
threads have unregistered it.
Change-Id: Ied970b0e9a9de512cec9e0841a591cc4d98e327d
Signed-off-by: Austin Hu <austin.hu@intel.com>
|
|
Change-Id: Iafa651485763e88ac05134e5b817cd8365c03d51
|
|
BUG: 22602467
Change use mode driver fw interface for 3.12
Change-Id: I068f723a13bd88e9554a257899285a79fa607d1a
Signed-off-by: hding3 <haitao.ding@intel.com>
|
|
Bug: 22618392
Read YUV data align with decoder output.
This means UV data are stored at the offset of alight height instead of video height.
Change-Id: I80b54adc23096bc6eade86807db01c1fb1ec6c4c
Signed-off-by: hding3 <haitao.ding@intel.com>
|
|
Bug: 22618392
Remove tile stride reset to fix hevc frame drop issue
Change-Id: I82d14f82ac4a17772026557356c8a835cf71d208
Signed-off-by: Haitao Ding <haitao.ding@intel.com>
|
|
Correct gralloc usage together with correct lock area trigger cache
flushing correctly, to resolve corruption during resolution change.
The memset workaround for Bug 19197299 is also removed since we didn't
observe it on DDK-1.5.
Bug: 22183821
Change-Id: If1c112988a300b7aa9054127aa37cc2cf50d5371
Signed-off-by: Tianmi Chen <tianmi.chen@intel.com>
|
|
Bug: 19576398
Change-Id: If002e1443f42b5e364df86f0dca16548cc4d51d1
Signed-off-by: hding3 <haitao.ding@intel.com>
|
|
Bug: 21156664
Change-Id: I790b7f39cea94cababf50b09876f7c68ed742518
|
|
Bug: 20914944
Bug: 20302101
Bug: 19412176
Bug: 21265944
Change-Id: I5b427711e1699e4d64895bc6fa1665e9383d8623
Signed-off-by: Pradeep Radhakrishna <pradeep.radhakrishna@intel.com>
|
|
when ion is enabled as backend of gralloc.
video driver needs to pass down fd of graphics buffer
when creat surface from it.
Orig-Author: Randy Xu <randy.xu@intel.com>
Bug: 18731042
BZ: 230505
Change-Id: Ic85a04e1b2766764fb85ff98503022fb7be6b54f
Signed-off-by: mahongpe <hongpeng.ma@intel.com>
|
|
The definition of macroblock_offset in VASliceParameterBufferVP8
has been changed. It is the offset to the first bit of MB from
the first byte of slice data.
pass color info to hwc with payload
Change-Id: I74c62a1a029009b4f4019d6cb684c879f0c4c2a8
Signed-off-by: Haitao Ding <haitao.ding@intel.com>
|
|
clear gfx buffer before decoding.
Bug: 19197299
Change-Id: I83f09fed7d38fc6698749756427771119a51003b
Signed-off-by: mahongpe <hongpeng.ma@intel.com>
|
|
Now that the kernel is using the proper DSI connector type
value, we can stop using the TMP_DSI type.
Bug: 19118845
Change-Id: I700ebb5598ad6b8f6999a849056c49f6c65389a1
Signed-off-by: Sean Paul <seanpaul@chromium.org>
|
|
Now that the DSI & Virtual connector types have been swapped,
we can accept the new DSI type in addition to the temporary
value. This will allow us to start using the new DSI type in
kernel such that we can then remove the temp values altogether
Bug: 19118845
Change-Id: I8af72bb95bd2f272cd0ff63e0e575b3b9e25a4b6
Signed-off-by: Sean Paul <seanpaul@chromium.org>
|
|
For primary display connector type, only accept the TMP_DSI
connector type. This will allow us to change the primary DSI
value in the kernel. We'll switch back once this is done.
Bug: 19118845
Change-Id: I4ccc760538789de669b47aa3bf6036069df80d51
Signed-off-by: Sean Paul <seanpaul@chromium.org>
|
|
For primary display connector type, accept both MIPI or
TMP_DSI types.
Bug: 19118845
Change-Id: I25dc29a8f3ebaee1c46fb6464607ae07bf0883af
Signed-off-by: Sean Paul <seanpaul@chromium.org>
|
|
This is already defined in drm_mode.h and xf86drmMode.h.
Bug: 19118845
Change-Id: I5186a6d7c9a869f607e67f1f9db482b0ffaa6271
Signed-off-by: Sean Paul <seanpaul@chromium.org>
|
|
MPEG2 driver implementation
Change-Id: I43933536803788f31445e4eeb00d95e9dfa620c6
Signed-off-by: Yuanjun Huang <yuanjun.huang@intel.com>
|
|
* commit 'e2d5072c4f0604bdebae4d964ccd2f4f125b4600':
Fix "error: implicit declaration of function 'malloc'".
|
|
Change-Id: I443e4ad88bf538406ddbb6fe4ba3e54eba34a995
|
|
* commit '4bac5bf335c63740958c98bab11b94dc37def621':
Fix "error: implicit declaration of function 'realloc'".
|
|
* commit '6e651afc8c61e7cc2a1f08078e908df2810af49a':
Fix "error: implicit declaration of function 'calloc'".
|
|
Change-Id: Ie37e0dc21b975f95c5c5600907641e662b3a51b2
|
|
Change-Id: I9692f3b660b4832605fcb03bc60f1572b47e6ca3
|
|
potential CTS issue such as: android.media.cts.EncodeDecodeTest#testVP8EncodeDecodeVideoFromBufferToSurfaceQVGA
* commit 'cd73c9d1a30131aecaf31eba26456e55f06d75a4':
psb-video: Change the surface type to be uncached. to fix potential CTS issue such as: android.media.cts.EncodeDecodeTest#testVP8EncodeDecodeVideoFromBufferToSurfaceQVGA
|
|
to fix potential CTS issue such as:
android.media.cts.EncodeDecodeTest#testVP8EncodeDecodeVideoFromBufferToSurfaceQVGA
Bug: 19183778
Bug: 19184881
Change-Id: Ifd7a4ccafeaec5fe7b84f18b4417e95822dc851f
Signed-off-by: Yuanjun Huang <yuanjun.huang@intel.com>
|
|
* commit '5526ee6314de2723fa8f651dec2ff337e4c7a7da':
<string.h>? No, <strings.h>.
|
|
Who still uses bzero anyway?
Change-Id: Ia6f3c406c18b01c44881686b807b04f9bf62abe5
|
|
* commit '7d8048d2853335f06ddd71539830cfc2a270585e':
Add missing <string.h> include.
|
|
Change-Id: Ie6ba07b09a2f357370a2aa1195762e7b48bdc10d
|
|
* commit 'd640886ce44885e2a70f4d31187d17f4ce08d52d':
Fix share info initialization issue when enabling VPP/FRC
|
|
add a flag to indicate share_info initilization state
Bug: 18629797
BZ: 230329
Change-Id: Ifd452233049a26d00742f59e0af9f624295fbdcb
Signed-off-by: Jason Hu <jason.hu@intel.com>
|
|
* commit 'c8e9f406b7c13d6eb717f5f1359469ddad181ed9':
Suppress clang/llvm compiler warnings before upstream fix.
|
|
Examples:
* psb_drv_video.c:211:63: comparison of constant 32 with expression
of type 'VAProfile' is always false [-Werror,-Wtautological-constant-out-of-range-compare]
* tng_hostcode.c:3663:47: implicit conversion from enumeration type 'enum _frame_template_type_'
to different enumeration type 'IMG_FRAME_TYPE' (aka 'enum * _frame_type_') [-Werror,-Wenum-conversion]
Change-Id: Iaa3fc2dac51ab6209d0047bd6d3c05b36cce6ea9
|
|
* commit 'fe436fee778fb0470d1d07150a8150eb38232533':
Fix VP8 encoder realted CTS issue .
|
|
Bug: 18670741
Change-Id: If2e3c934a575a90583c53db3f9d734c2998558fa
Signed-off-by: hding3 <haitao.ding@intel.com>
|
|
* commit 'b7f682774163606844fff8a0d6baf65a1158a206':
Suppress clang warning gnu-designator.
|
|
We could also change the source code to use standard syntax.
BUG: 18193625
Change-Id: I827a77f37907def68b34587ecc1447bf9ab88fd6
|
|
Bug: 18156738
BZ: 227732
Change-Id: Id2f135e3cf27a15300e19ccadd18ef257f13109f
Signed-off-by: Yuanjun Huang <yuanjun.huang@intel.com>
|
|
Remove ION memory manager due to regressions caused by its inclusion.
This reverts commit 7f1c4d0835202abcb6f7bf42a3a8813dc48a8e0c.
Bug: 18578655
Bug: 18527174
BZ: 230087
BZ: 229790
BZ: 229995
Change-Id: I56965a48f4fe24aff2317610153715c1a83d4ab5
Signed-off-by: Dale Stimson <dale.b.stimson@intel.com>
|
|
In current design, all YUV420 surface will be handled with VA_FOURCC_NV12.
The correct pixel format will be saved in extra_info[8].
VSP driver will indentify YV12 format with extra_info[8].
BZ: 229811
Bug: 17383204
Change-Id: I405b37bbac273e611b7415781b6dc586808b78ba
Signed-off-by: Wang Kun <kun.k.wang@intel.com>
|
|
Fix VED fw hang issue in Monkey test
Bug: 17656112
Bug: 18306960
Change-Id: Ia5fa51e02ccb2765c4308db603559181144efe07
Signed-off-by: hding3 <haitao.ding@intel.com>
|
|
Decrease isv log amount
BZ: 229745
Bug: 17383204
Change-Id: I5c09a56a6767ac1cc62a70ee2f37764fa98ed9eb
Signed-off-by: Xigui Wang <xigui.wang@intel.com>
|
|
when ion is enabled as backend of gralloc.
video driver needs to pass down fd of graphics buffer
when creat surface from it.
Orig-Author: Randy Xu <randy.xu@intel.com>
BZ: 224134
Bug: 17316531
Depends-on: Ife9513fe58e06ea2829987f328668f975a44b4e0
Depends-on: I4b8422d48b06419af6e48b631bacbe6c7169574e
Depends-on: Ia4c5bc80df952024a0973f2bacc42992301dfbf9
Depends-on: I172cdcb8cc86fd6fa3d3adadbab19b6fd3e6a18e
Depends-on: I2a2e721719349f755033c15bdeb732e7fb5c89ac
Depends-on: I48deabb107cc3abfa59eff5e012b51293fe6d74f
Depends-on: If2be266fc9063d7a2b701f85e8ace26aaf484ff5
Change-Id: I0ca0f06491e3fb498b3b41b70cbe1b6cf39e1918
Signed-off-by: mahongpe <hongpeng.ma@intel.com>
|
|
|
|
Bug: 18313857
Change-Id: Idfd1db35e780e94d79ec0b22f22a5e319c99e7c8
Signed-off-by: Thierry Strudel <thierry.strudel@intel.com>
|
|
1. builds VSP vpp files.
2. remove dependency on libvpp
3. Remove the limitation of va surface can't be NULL when vaCreateContext
4. Fix a multiple thread issue when vaCreateSurfaces
When try to invoke gralloc_lock for the same native handle from different
thread, will return lock err.
Add a interface lock to protect the access to native handle.
5. Remove the compiling warning for VPP/VP8.
Update the data structure for parameters
Add "unused" attribute for unused function parameter.
7. Add flag in vaCreateSurfaces2 to explicit set NOT initialized share info
when creating surfaces.
8. Copy VPP input share info to output
Bug: 17383204
BZ: 227971
Change-Id: I6cfe01494d12f8704145294d391f4241801b05f6
Signed-off-by: Xigui Wang <xigui.wang@intel.com>
Signed-off-by: pingshix <pingx.shi@intel.com>
|
|
Bug: 17656112
BZ: 226486
Change-Id: I2afd29356b600b3cc002bdd7394640122757a6ab
Signed-off-by: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
|
|
enhancement for maintaining the consistent of DBP also fix the CRC
issue in previous commit
Bug: 17693351
Signed-off-by: ywan171 <yi.a.wang@intel.com>
Change-Id: I354dc662085463472a328b680fa01abe01751916
|